To address our first objective, we estimated the age of the Norops humilis/quaggulus species complex and evaluated two potential dispersal patterns for this group. The
BEAST analysis yielded a mean crown group age of 17.2 Myr BP (range?=?14.2–20.6) for
the clade. We report this date cautiously because it was calculated from mitochondrial
data only. To evaluate the second objective, we used the ancestral range estimation
to evaluate dispersal patterns. Using the LAGRANGE output, we infer that the common
ancestor of the N. humilis/quaggulus complex originated in Panama before dispersing west to Costa Rica, and then north
into Nicaragua and Honduras. The south-to-north pattern and origin in Panama suggests
isolation of the ancestor of the N. humilis/quaggulus complex in the Talamancan region of extreme LCA. The divergence between Pacific and
Caribbean N. humilis in northern Costa Rica (Chorotega Block) was estimated to be 6.6 Myr BP (range?=?4.12–6.94),
which corresponds to the estimate of 5.4 Myr BP for the rise of the lower Central
American highlands 57]. This uplift represents a potential vicariant event responsible for the separation
of the Pacific (lineage 3) and Caribbean lineages (N. humilis lineages 4 and 5?+?N. quaggulus). However, the Tilarán range in northern Costa Rica is estimated to have originated
around 2 Myr BP 57], after the present split between Caribbean and Pacific lineages. Sister species are
often found on opposite sides of the Central American Highlands 20], 21], and while Caribbean lineages of N. humilis form a paraphyletic group, the fact that one clade is separated by the continental
divide is unsurprising. The ancestral area estimation suggests that early ancestors
of some of the lineages were present on both sides of the continental divide in Costa
Rica and Panama, corresponding to dates prior to the rise of the Central American
Highlands.
As in many other taxa (58] and sources within), LCA has served as a region of diversification for the Norops humilis/quaggulus complex (once again, distinct from the N. humilis species group). The northward distribution observed here is similar to that found
in eleutherodactyline (genera Craugastor19]; Pristimantis59]) and hylid frogs (Dendropsophus60]) that originated in South America and dispersed to Central America prior to the most
recent completion of the Isthmus of Panama (3–4 Myr BP). This is also similar to findings
in other groups of squamates 2], although it remains to be seen if this pattern of dispersal is shared by other lizard
species.
Our third objective was to examine polyphyly described by others 35], 36] of the N. humilis species group. The lack of support for a monophyletic N. humilis species group is consistent with results from Nicholson 35] and Poe 36], further illustrating the extent of convergence in morphological characters of mainland
anoles. All analyses agreed in assigning several species of the N. humilis species group into separate areas of the tree. Norops marsupialis, N. tropidonotus, and N. uniformis were placed in areas of the tree distant to a monophyletic N. humilis/quaggulus complex. Initially N. marsupialis was included in this study as a member of the N. humilis/quaggulus complex. Taylor 61] described N. marsupialis as a subspecies of N. humilis in 1956, and it was not elevated to specific status until 2015 34]. All of our analyses (Bayesian and ML for each gene region and combined) found support
for the specific status of N. marsupialis, and agreed to its placement as sister to a clade containing N. aquaticus and N. woodi (approx. 30 Myr divergent from the N. humilis/quaggulus complex, Fig. 6). Therefore, we confirm N. marsupialis as a distinct species morphologically similar to, but evolutionary distant from N. humilis/quaggulus, in concordance with with Köhler et al. 34]. Morphological convergence within anoles has long been reported, particularly for
Caribbean species 27], 62]–66] and is further demonstrated by our analysis of the N. humilis species group. Inclusion of the other three members of the group (N. compressicauda, N. notopholis and N. wampuensis) into the molecular phylogeny of Norops may yield further insight on their placement in the phylogeny, but their inclusion
is not likely to significantly alter the results obtained here. Additional work within
N. tropidonotus may also be necessary to examine the evolution that has occurred in that species,
as it occupies a broad geographic range (Köhler 2008 30]). Our data suggest that N. tropidonotus contains at least three deep mitochondrial divergences, and could potentially represent
multiple cryptic lineages. While N. quaggulus is nested within N. humilis, we do not recommend synonymizing the two species until further work is done to investigate
this complex. The deep divergences seen among the N. humilis clades may correspond to cryptic species, and if further work confirms the presence
of two or more species, the name N. quaggulus would have priority as the senior synonym.
The SAMOVA results indicate that limited mitochondrial gene flow is occurring among
localities, suggesting 12 distinct genetic groupings among our sample localities.
Therefore, we conclude that genetic differentiation within the N. humilis/quaggulus complex is significant enough to conclude that (1) population fragmentation has occurred
and (2) the complex does not represent a single panmictic population. The isolation,
coupled with the high genetic variance, supports all major lineages identified in
the Bayesian analyses as being distinct from one another, as well as further subdivision
within most lineages. While we do not suggest that the 12 groups correspond to separate
species, the presence of at least six well supported, divergent clades within the
N. humilis/quaggulus complex demonstrates that deep mitochondrial divergence has occurred within this
group, although nuclear evolution appears to be more conserved. This result is similar
to several studies on Caribbean anoles, which also display considerable mitochondrial
differentiation, with much less diversity in the nuclear DNA 67]–70]. This pronounced phylogeographic substructuring may be explained in part, by low
vagility in lizards 9]. Deep mitochondrial divergences coupled with conserved nuclear evolution as seen
here, may have at least two implications 1) ITS is more slowly evolving than the mitochondrial
genome of anoles, which we consider to be a likely scenario given that nuclear DNA
is generally regarded as experiencing slower rates of evolution 71]–73] and 2). The mitochondrial-nuclear relationships observed here are characteristic
of male-biased dispersal 74], 75], indicating that female N. humilis are more philopatric than males as seen in Caribbean anoles 76], 77].
The novel biogeographic pattern for Central American anoles revealed here illustrates
a need for further work on mainland Norops. What remains to be tested is whether the south-to-north dispersal route seen in
the N. humilis/quaggulus complex is repeated in other Norops groups. In addition, it is important to clarify where the Central American Norops lineage originated, how it dispersed throughout the mainland, and when these events
took place. Investigating the phylogeography of other widespread anoles may be highly
informative towards understanding other distribution patterns of the Central American
herpetofauna. There are several widespread Norops species and species complexes that vary in their ecological roles with corresponding
morphological features. These are grouped into designations called ecomodes, which
are distinct from ecomorphs, to accommodate mainland anoles (see 35]). Such an assortment of ecologically diverse anoles may provide good models for testing
the biogeographic hypotheses discussed here. Further studies on mainland Norops species are needed to test if the cryptic diversity suggested here is present in
other widespread species complexes within the genus.